Editorial illustration for ChatGPT Sketch Transforms Doodles Into Detailed AI Images
ChatGPT Sketch Turns Doodles Into AI Images
OpenAI rolled out ChatGPT Images 2.5 on Tuesday, and the update comes with a feature that lets you draw your way into a prompt instead of just typing one. Type @Sketch into the chat box, and a small canvas pops up where you can scribble whatever you're trying to describe. Tell ChatGPT what to do with that doodle, and it builds a finished image around it.
The pitch is simple: most people can't describe a pose, a layout, or a composition in words as well as they could just draw it, even badly. Sketch is meant to close that gap, turning a rough mouse-drawn shape into a starting point for a real image rather than making users hunt for the right adjectives.
OpenAI is also adding inline comments, so you can mark up a specific part of a generated image, a color, an object, a detail, and ask for a targeted fix without redoing the whole prompt. Alongside those additions, OpenAI says Images 2.5 improves lighting and texture quality, handles multi-step edits more reliably, and cuts generation latency by up to 50% compared to Images 2.0. It's live now for ChatGPT, ChatGPT Work, and Codex users on desktop, mobile, and web.
OpenAI announced ChatGPT Images 2.5 on Tuesday and is adding a new way to tell ChatGPT what you want it to make an image of: by drawing a doodle. With a new feature called Sketch, you can just draw something right inside ChatGPT and then tell ChatGPT how you want it to make an image based on that sketch.
Why this matters Sketch is a small feature with a bigger point: OpenAI is betting that typing a prompt isn't the fastest way to tell a model what's in your head, and for a lot of visual work, they're right. A rough circle with an arrow pointing at it says something a paragraph of adjectives can't. For developers building on top of image generation, this is a signal about where the interface is headed, less pure text, more mixed input like sketches, annotations, and reference points fed into the same request. That's worth watching if you're designing your own tools around the API rather than the ChatGPT app itself.
For founders, the interesting question is adoption, not novelty. Doodle-to-image isn't new, but bundling it directly into ChatGPT's existing image workflow removes a step most competitors still require. Whether people actually use it beyond demos is the thing to track over the next few months, since OpenAI has a habit of shipping features that look clever in a launch post and then quietly stall out in daily use.
Related Reading
Browse all OpenAI news →
LLMs & Generative AI ChatGPT's 'Nerdy' tweak rewards goblin metaphors in answers, study finds
LLMs & Generative AI Google tests visual 'magazine-style' UI for Gemini 3 Pro users
LLMs & Generative AI AI Engineers Face Rising Costs, Need New Strategies for Efficiency
Related Topic OpenAI's GPT Transcribe Cuts Error Rate to 3.31%, Improving on GPT-4o
Open Source
Related Topic OpenAI Calls Apple's Lawsuit 'Rotten to Its Core' in Rebuttal
Policy & Regulation
From Archives GPT-6 Astra Beats Portal Start to Finish Autonomously in a Day
From Archives ChatGPT regains 55.5% web traffic share as Gemini's surge fades
From Archives NYU Mathematician Accuses OpenAI of Unethical Tactics on Math Problem
Business & Startups
From Archives Anthropic's USD 517 Billion Compute Deals Trail OpenAI's 2030 Goal
Business & Startups
Common Questions Answered
How does the ChatGPT Sketch feature work in ChatGPT Images 2.5?
The Sketch feature allows users to type @Sketch into the chat box to open a canvas where they can draw or scribble their ideas. After creating a doodle, users can tell ChatGPT how they want it to transform that sketch into a finished, detailed image based on their instructions.
Why is ChatGPT Sketch better than typing text prompts for image generation?
Most people can communicate visual concepts like poses, layouts, and compositions more effectively through drawing than through written descriptions, even if the sketch is rough or poorly drawn. A simple sketch with visual elements like shapes and arrows can convey compositional intent that would require lengthy paragraphs of adjectives to describe in text form.
What does the Sketch feature signal about the future direction of AI image generation interfaces?
OpenAI is indicating that pure text prompts are not the most efficient way to communicate with AI models for visual work, and future interfaces will likely incorporate mixed input methods. This includes sketches, annotations, and reference points rather than relying solely on typed text descriptions.
When was the ChatGPT Sketch feature released?
OpenAI rolled out ChatGPT Images 2.5 with the Sketch feature on Tuesday, introducing this new way for users to create image prompts through drawing rather than typing.
Further Reading
- Exclusive: Hands on with ChatGPT's new image editor - Axios
- OpenAI launches ChatGPT Images 2.5, which it says cuts image generation latency by up to 50% vs. Images 2.0, and adds a Sketch feature for drawing in ChatGPT - Techmeme
- OpenAI's ChatGPT Images 2.5 Cuts Generation Time by 50% ... - AlphaSignal
- OpenAI, ChatGPT 이미지 2.5 출시... 생성 속도 향상 - Investing.com
- The new ChatGPT Images is here - OpenAI
